Defang Blog
May 2025 Defang Compose Update

May was a big month at Defang. We shipped support for managed LLMs in Playground, added MongoDB support on AWS, improved the Defang MCP Server, and dropped new AI samples to make deploying faster than ever.
🚀 Managed LLMs in Playground
You can now try managed LLMs directly in the Defang Playground. Defang makes it easy to use cloud-native language models across providers — and now you can test them instantly in the Playground.
- Managed LLM support
- Playground-ready
- Available in CLI v1.1.22 or higher
To use managed language models in your own Defang services, just add x-defang-llm: true — Defang will configure the appropriate roles and permissions for you.
Already built on the OpenAI API? No need to rewrite anything.
With Defang's OpenAI Access Gateway, you can run your existing apps on Claude, DeepSeek, Mistral, and more — using the same OpenAI format.
📦 MongoDB Preview on AWS
Last month, we added support for MongoDB-compatible workloads on AWS via Amazon DocumentDB.
Just add this to your compose.yaml:
Once you add x-defang-mongodb: true, Defang will auto-spin a DocumentDB cluster in your AWS — no setup needed.
🛠 MCP Server Improvements
We've made the MCP Server and CLI easier to use and deploy:
- Users are now prompted to agree to Terms of Service via the portal login
- MCP Server and CLI are now containerized, enabling faster setup, smoother deployments, and better portability across environments
🌎 Events and Community
We kicked off the month by sponsoring Vancouver's first Vibe Coding IRL Sprint. Jordan Stephens from Defang ran a hands-on workshop on "Ship AI Faster with Vertex AI" with GDG Vancouver (GDG Vancouver). Around the same time, our CTO and Co-founder Lio joined the GenAI Founders Fireside panel hosted by AInBC and AWS.
Big moment for the team — we won the Best Canadian Cloud Award at the Vancouver Cloud Summit. Right after, we hit the expo floor at Web Summit Vancouver as part of the BETA startup program and got featured by FoundersBeta as one of the Top 16 Startups to Watch.
Our Campus Advocates also kept the momentum going, hosting Defang events around the world with live demos and workshops.
Last month's Defang Coffee Chat brought together the community for product updates, live demos, and a great convo on vibe deploying.
We're back again on June 25 at 10 AM PST. Save your spot here.
We can't wait to see what you deploy with Defang. Join our Discord to ask questions, get support, and share your builds.
More coming in June.
Related posts

FastAPI deployment Heroku vs AWS / GCP with Defang
Defang brings Heroku’s simplicity to your own cloud: deploy FastAPI apps to AWS or GCP with a single command using your existing Docker Compose. You keep full ownership, pay raw cloud costs, and stay portable. No platform lock-in, same Heroku-style workflow.

September 2025 Defang Compose Update
September focused on refinement and speed. We tightened deployment workflows, strengthened authentication, and made collaboration smoother across every environment. We improved Portal stability, added Railpack build logs, and refined Compose handling for managed services like Postgres, Mongo, and Redis. WAF is now enabled for high availability on AWS, and our VS Code extension is verified on the Marketplace with an easier MCP setup built right into the installer. Here’s what’s new.

n8n is blowing up. Here's how to deploy it to AWS in minutes.
Learn about n8n, how it's revolutionizing AI, and how to deploy it securely to your AWS account in 5min.